md: fix deadlock error in recent patch.

A recent patch aimed to cause md_write_start() to fail (rather than
block) when the mddev was suspending, so as to avoid deadlocks.
Unfortunately the test in wait_event() was wrong, and it didn't change
behaviour at all.

We wait_event() must wait until the metadata is written OR the array is
suspending.

Fixes: cc27b0c78c ("md: fix deadlock between mddev_suspend() and md_write_start()")
Cc: stable@vger.kernel.org
Reported-by: Xiao Ni <xni@redhat.com>
Signed-off-by: NeilBrown <neilb@suse.com>
Signed-off-by: Shaohua Li <shli@fb.com>
This commit is contained in:
NeilBrown 2017-10-05 16:23:16 +11:00 committed by Shaohua Li
parent 9e66317d3c
commit d47c8ad261
1 changed files with 2 additions and 1 deletions

View File

@ -8039,7 +8039,8 @@ bool md_write_start(struct mddev *mddev, struct bio *bi)
if (did_change) if (did_change)
sysfs_notify_dirent_safe(mddev->sysfs_state); sysfs_notify_dirent_safe(mddev->sysfs_state);
wait_event(mddev->sb_wait, wait_event(mddev->sb_wait,
!test_bit(MD_SB_CHANGE_PENDING, &mddev->sb_flags) && !mddev->suspended); !test_bit(MD_SB_CHANGE_PENDING, &mddev->sb_flags) ||
mddev->suspended);
if (test_bit(MD_SB_CHANGE_PENDING, &mddev->sb_flags)) { if (test_bit(MD_SB_CHANGE_PENDING, &mddev->sb_flags)) {
percpu_ref_put(&mddev->writes_pending); percpu_ref_put(&mddev->writes_pending);
return false; return false;